Over the last decade, the amount of Arabic content created on websites and social media has grown significantly. Opinions are shared openly and freely on social media and thus provide a rich source for trend analyses, which are accomplished by conventional methods of language interpretation, such as sentiment analysis. Due to its accuracy in studying unstructured data, deep learning has been increasingly used to test opinions. Recurrent neural networks (RNNs) are a promising approach in textual analysis and exhibit large morphological variations. In total, 193 studies used RNNs in English-language sentiment analysis, and 24 studies used RNNs in Arabic-language sentiment analysis. Those studies varied in the areas they address, the functionality and weaknesses of the models, and the number and scale of the available datasets for different dialects. Such variations are worthy of attention and monitoring; thus, this paper presents a systematic examination of the literature to label, evaluate, and identify state-of-the-art studies using RNNs for Arabic sentiment analysis.

在过去的十年中，在网站和社交媒体上创建的阿拉伯语内容数量显着增长。意见在社交媒体上自由开放地共享，从而为趋势分析提供了丰富的资源，这些趋势分析是通过传统的语言解释方法（例如情感分析）来完成的。由于其在研究非结构化数据方面的准确性，因此深度学习已越来越多地用于测试观点。递归神经网络（RNN）在文本分析中是一种很有前途的方法，并且表现出较大的形态变化。总共193项研究在英语情感分析中使用了RNN，而24项研究在阿拉伯语情感分析中使用了RNN。这些研究在其解决的领域，模型的功能和弱点以及针对不同方言的可用数据集的数量和规模方面各不相同。这种变化值得关注和监测；因此，本文对使用RNN进行阿拉伯语情感分析的文献进行系统地检查，以标记，评估和识别最新研究。